[
https://issues.apache.org/jira/browse/SPARK-46516?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Guram Savinov updated SPARK-46516:
----------------------------------
Description:
>From the docs: spark.sql.autoBroadcastJoinThreshold - Configures the maximum
>size in bytes for a table that will be broadcasted to all worker nodes when
>performing a join.
[https://spark.apache.org/docs/3.5.0/configuration.html#runtime-sql-configuration]
In fact Spark compares plan.statistics.sizeInBytes of a project (columns
selected in join), not a relation size.
[https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/joins.scala#L368]
Join can select only a few columns and sizeInBytes will be lesser than
autoBroadcastJoinThreshold, but broadcasted table can be huge and it is loaded
entirely into drivers memory which can lead to OOM.
spark.sql.autoBroadcastJoinThreshold parameter compared to projection instead
of broadcasted table size seems quite risky feature: there will be more
broadcasted relations but more chances to get OOM on the driver too.
Original task and test when autobroadcast compared to relation totalSize:
https://issues.apache.org/jira/browse/SPARK-2393
[https://github.com/apache/spark/pull/1238/files#diff-00485e6cae519f81adca5ceee66227c6eae35db709619d505468f8765175ac18R39]
Task and PR where autoBroadcastJoinThreshold started to be compared to project
of a plan instead of relations:
https://issues.apache.org/jira/browse/SPARK-13329
[https://github.com/apache/spark/pull/11210]
Related topic on SO:
[https://stackoverflow.com/questions/74435020/how-dataframe-count-selects-broadcasthashjoin-while-dataframe-show-selects-s]
was:
>From the docs: spark.sql.autoBroadcastJoinThreshold - Configures the maximum
>size in bytes for a table that will be broadcasted to all worker nodes when
>performing a join.
[https://spark.apache.org/docs/3.5.0/configuration.html#runtime-sql-configuration]
In fact Spark compares plan.statistics.sizeInBytes of a project (columns
selected in join), not a relation size.
[https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/joins.scala#L368]
Join can select only a few columns and sizeInBytes will be lesser than
autoBroadcastJoinThreshold, but broadcasted table can be huge and it is loaded
entirely into drivers memory which can lead to OOM.
spark.sql.autoBroadcastJoinThreshold parameter seems useless when its not
compared to broadcasted table size.
Original task and test when autobroadcast compared to relation totalSize:
https://issues.apache.org/jira/browse/SPARK-2393
[https://github.com/apache/spark/pull/1238/files#diff-00485e6cae519f81adca5ceee66227c6eae35db709619d505468f8765175ac18R39]
Task and PR where autoBroadcastJoinThreshold started to be compared to project
of a plan instead of relations:
https://issues.apache.org/jira/browse/SPARK-13329
[https://github.com/apache/spark/pull/11210]
Related topic on SO:
[https://stackoverflow.com/questions/74435020/how-dataframe-count-selects-broadcasthashjoin-while-dataframe-show-selects-s]
> autoBroadcastJoinThreshold compared to project of a plan not a relation size
> ----------------------------------------------------------------------------
>
> Key: SPARK-46516
> URL: https://issues.apache.org/jira/browse/SPARK-46516
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 3.1.1
> Reporter: Guram Savinov
> Priority: Major
>
> From the docs: spark.sql.autoBroadcastJoinThreshold - Configures the maximum
> size in bytes for a table that will be broadcasted to all worker nodes when
> performing a join.
> [https://spark.apache.org/docs/3.5.0/configuration.html#runtime-sql-configuration]
> In fact Spark compares plan.statistics.sizeInBytes of a project (columns
> selected in join), not a relation size.
> [https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/joins.scala#L368]
> Join can select only a few columns and sizeInBytes will be lesser than
> autoBroadcastJoinThreshold, but broadcasted table can be huge and it is
> loaded entirely into drivers memory which can lead to OOM.
> spark.sql.autoBroadcastJoinThreshold parameter compared to projection instead
> of broadcasted table size seems quite risky feature: there will be more
> broadcasted relations but more chances to get OOM on the driver too.
>
> Original task and test when autobroadcast compared to relation totalSize:
> https://issues.apache.org/jira/browse/SPARK-2393
> [https://github.com/apache/spark/pull/1238/files#diff-00485e6cae519f81adca5ceee66227c6eae35db709619d505468f8765175ac18R39]
>
> Task and PR where autoBroadcastJoinThreshold started to be compared to
> project of a plan instead of relations:
> https://issues.apache.org/jira/browse/SPARK-13329
> [https://github.com/apache/spark/pull/11210]
>
> Related topic on SO:
> [https://stackoverflow.com/questions/74435020/how-dataframe-count-selects-broadcasthashjoin-while-dataframe-show-selects-s]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
